python 的 OpenCV 3.2，Windows 中的 CUDA
OpenCV 3.2 for python with CUDA in Windows
我用 cmake 和 Visual Studio 构建了 opencv3.2 12 (2013) CUDA 支持和 python3.5。所有库都已成功构建，我可以使用 python.
运行 opencv 命令
这是命令的输出 print(cv2.getBuildInformation())
General configuration for OpenCV 3.2.0 =====================================
Version control: unknown
Extra modules:
Location (extra): D:/OpenCV/opencv_contrib/modules
Version control (extra): unknown
Platform:
Timestamp: 2017-06-09T14:07:41Z
Host: Windows 10.0.14393 AMD64
CMake: 3.9.0-rc2
CMake generator: Visual Studio 12 2013 Win64
CMake build tool: C:/Program Files (x86)/MSBuild/12.0/bin/MSBuild.exe
MSVC: 1800
C/C++:
Built as dynamic libs?: YES
C++ Compiler: D:/Program Files (x86)/Microsoft Visual Studio 12.0/VC/bin/x86_amd64/cl.exe (ver 18.0.21005.1)
C++ flags (Release): /DWIN32 /D_WINDOWS /W4 /GR /EHa /D _CRT_SECURE_NO_DEPRECATE /D _CRT_NONSTDC_NO_DEPRECATE /D _SCL_SECURE_NO_WARNINGS /Gy /bigobj /arch:AVX2 /Oi /wd4127 /wd4251 /wd4324 /wd4275 /wd4589 /MP4 /MD /O2 /Ob2 /DNDEBUG /Zi
C++ flags (Debug): /DWIN32 /D_WINDOWS /W4 /GR /EHa /D _CRT_SECURE_NO_DEPRECATE /D _CRT_NONSTDC_NO_DEPRECATE /D _SCL_SECURE_NO_WARNINGS /Gy /bigobj /arch:AVX2 /Oi /wd4127 /wd4251 /wd4324 /wd4275 /wd4589 /MP4 /MDd /Zi /Ob0 /Od /RTC1
C Compiler: D:/Program Files (x86)/Microsoft Visual Studio 12.0/VC/bin/x86_amd64/cl.exe
C flags (Release): /DWIN32 /D_WINDOWS /W3 /D _CRT_SECURE_NO_DEPRECATE /D _CRT_NONSTDC_NO_DEPRECATE /D _SCL_SECURE_NO_WARNINGS /Gy /bigobj /arch:AVX2 /Oi /MP4 /MD /O2 /Ob2 /DNDEBUG /Zi
C flags (Debug): /DWIN32 /D_WINDOWS /W3 /D _CRT_SECURE_NO_DEPRECATE /D _CRT_NONSTDC_NO_DEPRECATE /D _SCL_SECURE_NO_WARNINGS /Gy /bigobj /arch:AVX2 /Oi /MP4 /MDd /Zi /Ob0 /Od /RTC1
Linker flags (Release): /machine:x64 /INCREMENTAL:NO /debug
Linker flags (Debug): /machine:x64 /debug /INCREMENTAL
ccache: NO
Precompiled headers: YES
Extra dependencies: comctl32 gdi32 ole32 setupapi ws2_32 vfw32 C:/Users/Ramesh/Anaconda3/Library/lib/hdf5.lib cudart nppc nppi npps cublas cufft -LIBPATH:C:/Program Files/NVIDIA GPU Computing Toolkit/CUDA/v8.0/lib/x64
3rdparty dependencies: zlib libjpeg libwebp libpng libtiff libjasper IlmImf libprotobuf
OpenCV modules:
To be built: cudev core cudaarithm flann hdf imgproc ml reg surface_matching video cudabgsegm cudafilters cudaimgproc cudawarping dnn fuzzy imgcodecs photo shape videoio cudacodec highgui objdetect plot ts xobjdetect xphoto bgsegm bioinspired dpm face features2d line_descriptor saliency text calib3d ccalib cudafeatures2d cudalegacy cudaobjdetect cudaoptflow cudastereo datasets rgbd stereo superres tracking videostab xfeatures2d ximgproc aruco optflow phase_unwrapping stitching structured_light python3
Disabled: world contrib_world
Disabled by dependency: -
Unavailable: java python2 viz cnn_3dobj cvv freetype matlab sfm
Windows RT support: NO
GUI:
QT: NO
Win32 UI: YES
OpenGL support: NO
VTK support: NO
Media I/O:
ZLib: build (ver 1.2.8)
JPEG: build (ver 90)
WEBP: build (ver 0.3.1)
PNG: build (ver 1.6.24)
TIFF: build (ver 42 - 4.0.2)
JPEG 2000: build (ver 1.900.1)
OpenEXR: build (ver 1.7.1)
GDAL: NO
GDCM: NO
Video I/O:
Video for Windows: YES
DC1394 1.x: NO
DC1394 2.x: NO
FFMPEG: YES (prebuilt binaries)
avcodec: YES (ver 57.64.101)
avformat: YES (ver 57.56.100)
avutil: YES (ver 55.34.100)
swscale: YES (ver 4.2.100)
avresample: YES (ver 3.1.0)
GStreamer: NO
OpenNI: NO
OpenNI PrimeSensor Modules: NO
OpenNI2: NO
PvAPI: NO
GigEVisionSDK: NO
DirectShow: YES
Media Foundation: NO
XIMEA: NO
Intel PerC: NO
Parallel framework: Concurrency
Other third-party libraries:
Use IPP: 9.0.1 [9.0.1]
at: D:/OpenCV/build2/3rdparty/ippicv/ippicv_win
Use IPP Async: NO
Use Lapack: NO
Use Eigen: YES (ver 3.3.3)
Use Cuda: YES (ver 8.0)
Use OpenCL: YES
Use OpenVX: NO
Use custom HAL: NO
NVIDIA CUDA
Use CUFFT: YES
Use CUBLAS: YES
USE NVCUVID: NO
NVIDIA GPU arch: 20 30 35 37 50 52 60 61
NVIDIA PTX archs:
Use fast math: YES
OpenCL: <Dynamic loading of OpenCL library>
Include path: D:/OpenCV/sources/3rdparty/include/opencl/1.2
Use AMDFFT: NO
Use AMDBLAS: NO
Python 2:
Interpreter: NO
Python 3:
Interpreter: C:/Users/Ramesh/Anaconda3/python.exe (ver 3.5.3)
Libraries: C:/Users/Ramesh/Anaconda3/libs/python35.lib (ver 3.5.3)
numpy: C:/Users/Ramesh/Anaconda3/lib/site-packages/numpy/core/include (ver 1.13.0)
packages path: C:/Users/Ramesh/Anaconda3/Lib/site-packages
Python (for build): C:/Users/Ramesh/Anaconda3/python.exe
Java:
ant: NO
JNI: C:/Program Files/Java/jdk1.8.0_112/include C:/Program Files/Java/jdk1.8.0_112/include/win32 C:/Program Files/Java/jdk1.8.0_112/include
Java wrappers: NO
Java tests: NO
Matlab: Matlab not found or implicitly disabled
Documentation:
Doxygen: D:/Program Files/doxygen/bin/doxygen.exe (ver 1.8.13)
Tests and samples:
Tests: NO
Performance tests: NO
C/C++ Examples: NO
Install path: D:/OpenCV/build2/install
cvconfig.h is in: D:/OpenCV/build2
-----------------------------------------------------------------
但是我不能使用 cv2.cuda 包中的函数。当我运行help(cv2.cuda)
Help on module cv2.cuda in cv2:
NAME
cv2.cuda
DATA
SURF_CUDA_ANGLE_ROW = 5
SURF_CUDA_HESSIAN_ROW = 6
SURF_CUDA_LAPLACIAN_ROW = 2
SURF_CUDA_OCTAVE_ROW = 3
SURF_CUDA_ROWS_COUNT = 7
SURF_CUDA_SIZE_ROW = 4
SURF_CUDA_X_ROW = 0
SURF_CUDA_Y_ROW = 1
FILE
(built-in)
结果说模块中没有函数cv2.cuda
我的问题是，
- 有没有办法编译
opencv3.2 并 CUDA 支持 python3.5..？正如 this 问题中所述，在 opencv 2.4.7 中是不可能的
- 如果没有办法，我可以将
c++函数绑定到python..吗？我经历了 opencv doc，我无法理解如何将新模块扩展到 python。当我 google 它时，我得到了一些关于编写包装器的答案（所有答案都超过 4 年）但是我不明白如何在 2 种语言之间绑定数据结构，比如 c++ Mat 和 numpy.ndarray.
任何帮助将不胜感激！！
我能够按照 here 中的教程找到获得 cuda 对 OpenCV python 支持的方法（我没有对此进行测试）。据我所知 GPUMat 不可用。但是 UMat 可用于 python 可用于访问 OpenCL.
如 答案中所述，UMat 是
A unified abstraction cv::UMat that enables the same APIs to be implemented using CPU or OpenCL code, without a requirement to call OpenCL accelerated version explicitly. These functions use an OpenCL-enabled GPU if exists in the system, and automatically switch to CPU operation otherwise.
如 OpenCL 部分 here 中所述，您只需将普通矩阵（在 python np.ndarrays 中）更改为 UMats（在python cv2.UMat) 以获得 OpenCL 支持。
但是当我需要 cuda 支持时我所做的是，我在 c++ 中将该部分写为函数并提供 cuda 支持。我使该函数以 cv::Mat 图像作为参数，并以 return cv::Mat 作为结果。
然后按照 here 所述为该函数编写一个 python 包装器。
那是有点旧的代码。但它对我有用 python 3.5 和 python 2.7 与 OpenCV 3.4
我用 cmake 和 Visual Studio 构建了 opencv3.2 12 (2013) CUDA 支持和 python3.5。所有库都已成功构建，我可以使用 python.
opencv 命令
这是命令的输出 print(cv2.getBuildInformation())
General configuration for OpenCV 3.2.0 =====================================
Version control: unknown
Extra modules:
Location (extra): D:/OpenCV/opencv_contrib/modules
Version control (extra): unknown
Platform:
Timestamp: 2017-06-09T14:07:41Z
Host: Windows 10.0.14393 AMD64
CMake: 3.9.0-rc2
CMake generator: Visual Studio 12 2013 Win64
CMake build tool: C:/Program Files (x86)/MSBuild/12.0/bin/MSBuild.exe
MSVC: 1800
C/C++:
Built as dynamic libs?: YES
C++ Compiler: D:/Program Files (x86)/Microsoft Visual Studio 12.0/VC/bin/x86_amd64/cl.exe (ver 18.0.21005.1)
C++ flags (Release): /DWIN32 /D_WINDOWS /W4 /GR /EHa /D _CRT_SECURE_NO_DEPRECATE /D _CRT_NONSTDC_NO_DEPRECATE /D _SCL_SECURE_NO_WARNINGS /Gy /bigobj /arch:AVX2 /Oi /wd4127 /wd4251 /wd4324 /wd4275 /wd4589 /MP4 /MD /O2 /Ob2 /DNDEBUG /Zi
C++ flags (Debug): /DWIN32 /D_WINDOWS /W4 /GR /EHa /D _CRT_SECURE_NO_DEPRECATE /D _CRT_NONSTDC_NO_DEPRECATE /D _SCL_SECURE_NO_WARNINGS /Gy /bigobj /arch:AVX2 /Oi /wd4127 /wd4251 /wd4324 /wd4275 /wd4589 /MP4 /MDd /Zi /Ob0 /Od /RTC1
C Compiler: D:/Program Files (x86)/Microsoft Visual Studio 12.0/VC/bin/x86_amd64/cl.exe
C flags (Release): /DWIN32 /D_WINDOWS /W3 /D _CRT_SECURE_NO_DEPRECATE /D _CRT_NONSTDC_NO_DEPRECATE /D _SCL_SECURE_NO_WARNINGS /Gy /bigobj /arch:AVX2 /Oi /MP4 /MD /O2 /Ob2 /DNDEBUG /Zi
C flags (Debug): /DWIN32 /D_WINDOWS /W3 /D _CRT_SECURE_NO_DEPRECATE /D _CRT_NONSTDC_NO_DEPRECATE /D _SCL_SECURE_NO_WARNINGS /Gy /bigobj /arch:AVX2 /Oi /MP4 /MDd /Zi /Ob0 /Od /RTC1
Linker flags (Release): /machine:x64 /INCREMENTAL:NO /debug
Linker flags (Debug): /machine:x64 /debug /INCREMENTAL
ccache: NO
Precompiled headers: YES
Extra dependencies: comctl32 gdi32 ole32 setupapi ws2_32 vfw32 C:/Users/Ramesh/Anaconda3/Library/lib/hdf5.lib cudart nppc nppi npps cublas cufft -LIBPATH:C:/Program Files/NVIDIA GPU Computing Toolkit/CUDA/v8.0/lib/x64
3rdparty dependencies: zlib libjpeg libwebp libpng libtiff libjasper IlmImf libprotobuf
OpenCV modules:
To be built: cudev core cudaarithm flann hdf imgproc ml reg surface_matching video cudabgsegm cudafilters cudaimgproc cudawarping dnn fuzzy imgcodecs photo shape videoio cudacodec highgui objdetect plot ts xobjdetect xphoto bgsegm bioinspired dpm face features2d line_descriptor saliency text calib3d ccalib cudafeatures2d cudalegacy cudaobjdetect cudaoptflow cudastereo datasets rgbd stereo superres tracking videostab xfeatures2d ximgproc aruco optflow phase_unwrapping stitching structured_light python3
Disabled: world contrib_world
Disabled by dependency: -
Unavailable: java python2 viz cnn_3dobj cvv freetype matlab sfm
Windows RT support: NO
GUI:
QT: NO
Win32 UI: YES
OpenGL support: NO
VTK support: NO
Media I/O:
ZLib: build (ver 1.2.8)
JPEG: build (ver 90)
WEBP: build (ver 0.3.1)
PNG: build (ver 1.6.24)
TIFF: build (ver 42 - 4.0.2)
JPEG 2000: build (ver 1.900.1)
OpenEXR: build (ver 1.7.1)
GDAL: NO
GDCM: NO
Video I/O:
Video for Windows: YES
DC1394 1.x: NO
DC1394 2.x: NO
FFMPEG: YES (prebuilt binaries)
avcodec: YES (ver 57.64.101)
avformat: YES (ver 57.56.100)
avutil: YES (ver 55.34.100)
swscale: YES (ver 4.2.100)
avresample: YES (ver 3.1.0)
GStreamer: NO
OpenNI: NO
OpenNI PrimeSensor Modules: NO
OpenNI2: NO
PvAPI: NO
GigEVisionSDK: NO
DirectShow: YES
Media Foundation: NO
XIMEA: NO
Intel PerC: NO
Parallel framework: Concurrency
Other third-party libraries:
Use IPP: 9.0.1 [9.0.1]
at: D:/OpenCV/build2/3rdparty/ippicv/ippicv_win
Use IPP Async: NO
Use Lapack: NO
Use Eigen: YES (ver 3.3.3)
Use Cuda: YES (ver 8.0)
Use OpenCL: YES
Use OpenVX: NO
Use custom HAL: NO
NVIDIA CUDA
Use CUFFT: YES
Use CUBLAS: YES
USE NVCUVID: NO
NVIDIA GPU arch: 20 30 35 37 50 52 60 61
NVIDIA PTX archs:
Use fast math: YES
OpenCL: <Dynamic loading of OpenCL library>
Include path: D:/OpenCV/sources/3rdparty/include/opencl/1.2
Use AMDFFT: NO
Use AMDBLAS: NO
Python 2:
Interpreter: NO
Python 3:
Interpreter: C:/Users/Ramesh/Anaconda3/python.exe (ver 3.5.3)
Libraries: C:/Users/Ramesh/Anaconda3/libs/python35.lib (ver 3.5.3)
numpy: C:/Users/Ramesh/Anaconda3/lib/site-packages/numpy/core/include (ver 1.13.0)
packages path: C:/Users/Ramesh/Anaconda3/Lib/site-packages
Python (for build): C:/Users/Ramesh/Anaconda3/python.exe
Java:
ant: NO
JNI: C:/Program Files/Java/jdk1.8.0_112/include C:/Program Files/Java/jdk1.8.0_112/include/win32 C:/Program Files/Java/jdk1.8.0_112/include
Java wrappers: NO
Java tests: NO
Matlab: Matlab not found or implicitly disabled
Documentation:
Doxygen: D:/Program Files/doxygen/bin/doxygen.exe (ver 1.8.13)
Tests and samples:
Tests: NO
Performance tests: NO
C/C++ Examples: NO
Install path: D:/OpenCV/build2/install
cvconfig.h is in: D:/OpenCV/build2
-----------------------------------------------------------------
但是我不能使用 cv2.cuda 包中的函数。当我运行help(cv2.cuda)
Help on module cv2.cuda in cv2:
NAME
cv2.cuda
DATA
SURF_CUDA_ANGLE_ROW = 5
SURF_CUDA_HESSIAN_ROW = 6
SURF_CUDA_LAPLACIAN_ROW = 2
SURF_CUDA_OCTAVE_ROW = 3
SURF_CUDA_ROWS_COUNT = 7
SURF_CUDA_SIZE_ROW = 4
SURF_CUDA_X_ROW = 0
SURF_CUDA_Y_ROW = 1
FILE
(built-in)
结果说模块中没有函数cv2.cuda
我的问题是，
- 有没有办法编译
opencv3.2并CUDA支持python3.5..？正如 this 问题中所述，在opencv 2.4.7 中是不可能的
- 如果没有办法，我可以将
c++函数绑定到python..吗？我经历了 opencv doc，我无法理解如何将新模块扩展到 python。当我 google 它时，我得到了一些关于编写包装器的答案（所有答案都超过 4 年）但是我不明白如何在 2 种语言之间绑定数据结构，比如c++ Mat和numpy.ndarray.
任何帮助将不胜感激！！
我能够按照 here 中的教程找到获得 cuda 对 OpenCV python 支持的方法（我没有对此进行测试）。据我所知 GPUMat 不可用。但是 UMat 可用于 python 可用于访问 OpenCL.
如 UMat 是
A unified abstraction cv::UMat that enables the same APIs to be implemented using CPU or OpenCL code, without a requirement to call OpenCL accelerated version explicitly. These functions use an OpenCL-enabled GPU if exists in the system, and automatically switch to CPU operation otherwise.
如 OpenCL 部分 here 中所述，您只需将普通矩阵（在 python np.ndarrays 中）更改为 UMats（在python cv2.UMat) 以获得 OpenCL 支持。
但是当我需要 cuda 支持时我所做的是，我在 c++ 中将该部分写为函数并提供 cuda 支持。我使该函数以 cv::Mat 图像作为参数，并以 return cv::Mat 作为结果。
然后按照 here 所述为该函数编写一个 python 包装器。
那是有点旧的代码。但它对我有用 python 3.5 和 python 2.7 与 OpenCV 3.4